CN112085830B - Optical coherence angiography imaging method based on machine learning - Google Patents
Optical coherence angiography imaging method based on machine learning Download PDFInfo
- Publication number
- CN112085830B CN112085830B CN201910513946.0A CN201910513946A CN112085830B CN 112085830 B CN112085830 B CN 112085830B CN 201910513946 A CN201910513946 A CN 201910513946A CN 112085830 B CN112085830 B CN 112085830B
- Authority
- CN
- China
- Prior art keywords
- machine learning
- network model
- oct
- training
- images
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
- 238000010801 machine learning Methods 0.000 title claims abstract description 70
- 238000003384 imaging method Methods 0.000 title claims abstract description 37
- 238000002583 angiography Methods 0.000 title claims abstract description 20
- 230000003287 optical effect Effects 0.000 title claims abstract description 20
- 238000012549 training Methods 0.000 claims abstract description 62
- FCKYPQBAHLOOJQ-UHFFFAOYSA-N Cyclohexane-1,2-diaminetetraacetic acid Chemical compound OC(=O)CN(CC(O)=O)C1CCCCC1N(CC(O)=O)CC(O)=O FCKYPQBAHLOOJQ-UHFFFAOYSA-N 0.000 claims abstract description 32
- 238000004422 calculation algorithm Methods 0.000 claims abstract description 22
- 238000000034 method Methods 0.000 claims abstract description 12
- 230000000694 effects Effects 0.000 claims abstract description 8
- 238000012360 testing method Methods 0.000 claims description 11
- 238000013528 artificial neural network Methods 0.000 claims description 7
- 238000013527 convolutional neural network Methods 0.000 claims description 7
- 238000011056 performance test Methods 0.000 claims description 7
- 230000003044 adaptive effect Effects 0.000 claims description 4
- 238000012216 screening Methods 0.000 claims description 4
- 238000005457 optimization Methods 0.000 claims description 3
- 230000008569 process Effects 0.000 claims description 3
- 230000008485 antagonism Effects 0.000 claims 1
- 230000000306 recurrent effect Effects 0.000 claims 1
- 230000002792 vascular Effects 0.000 abstract description 3
- 238000001514 detection method Methods 0.000 abstract description 2
- 238000012014 optical coherence tomography Methods 0.000 description 33
- 238000005516 engineering process Methods 0.000 description 10
- 238000013135 deep learning Methods 0.000 description 5
- 210000004204 blood vessel Anatomy 0.000 description 4
- 230000008859 change Effects 0.000 description 4
- 238000011161 development Methods 0.000 description 4
- 230000018109 developmental process Effects 0.000 description 4
- 206010044565 Tremor Diseases 0.000 description 3
- 230000001427 coherent effect Effects 0.000 description 3
- 230000006870 function Effects 0.000 description 3
- 238000012545 processing Methods 0.000 description 3
- 210000001525 retina Anatomy 0.000 description 3
- ORILYTVJVMAKLC-UHFFFAOYSA-N Adamantane Natural products C1C(C2)CC3CC1CC2C3 ORILYTVJVMAKLC-UHFFFAOYSA-N 0.000 description 2
- 238000013473 artificial intelligence Methods 0.000 description 2
- 238000004364 calculation method Methods 0.000 description 2
- 230000019771 cognition Effects 0.000 description 2
- 210000003743 erythrocyte Anatomy 0.000 description 2
- 238000011160 research Methods 0.000 description 2
- 230000029058 respiratory gaseous exchange Effects 0.000 description 2
- 230000035945 sensitivity Effects 0.000 description 2
- 238000003325 tomography Methods 0.000 description 2
- 244000025254 Cannabis sativa Species 0.000 description 1
- 235000012766 Cannabis sativa ssp. sativa var. sativa Nutrition 0.000 description 1
- 235000012765 Cannabis sativa ssp. sativa var. spontanea Nutrition 0.000 description 1
- 241000282412 Homo Species 0.000 description 1
- 230000004913 activation Effects 0.000 description 1
- 238000000149 argon plasma sintering Methods 0.000 description 1
- 230000006399 behavior Effects 0.000 description 1
- 230000017531 blood circulation Effects 0.000 description 1
- 235000009120 camo Nutrition 0.000 description 1
- 235000005607 chanvre indien Nutrition 0.000 description 1
- 238000012512 characterization method Methods 0.000 description 1
- 239000002872 contrast media Substances 0.000 description 1
- 239000008358 core component Substances 0.000 description 1
- 125000004122 cyclic group Chemical group 0.000 description 1
- 238000007418 data mining Methods 0.000 description 1
- 238000003745 diagnosis Methods 0.000 description 1
- 238000002474 experimental method Methods 0.000 description 1
- 239000011487 hemp Substances 0.000 description 1
- 238000000338 in vitro Methods 0.000 description 1
- 238000001727 in vivo Methods 0.000 description 1
- 238000011423 initialization method Methods 0.000 description 1
- 238000002372 labelling Methods 0.000 description 1
- 230000007774 longterm Effects 0.000 description 1
- 230000007246 mechanism Effects 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000010606 normalization Methods 0.000 description 1
- 230000035479 physiological effects, processes and functions Effects 0.000 description 1
- 238000007637 random forest analysis Methods 0.000 description 1
- 230000002787 reinforcement Effects 0.000 description 1
- 230000000717 retained effect Effects 0.000 description 1
- 230000011218 segmentation Effects 0.000 description 1
- 239000004065 semiconductor Substances 0.000 description 1
- 230000003068 static effect Effects 0.000 description 1
- 238000012706 support-vector machine Methods 0.000 description 1
- 230000009897 systematic effect Effects 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T17/00—Three dimensional [3D] modelling, e.g. data description of 3D objects
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/30—Determination of transform parameters for the alignment of images, i.e. image registration
- G06T7/33—Determination of transform parameters for the alignment of images, i.e. image registration using feature-based methods
- G06T7/337—Determination of transform parameters for the alignment of images, i.e. image registration using feature-based methods involving reference images or patches
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/30—Determination of transform parameters for the alignment of images, i.e. image registration
- G06T7/33—Determination of transform parameters for the alignment of images, i.e. image registration using feature-based methods
- G06T7/344—Determination of transform parameters for the alignment of images, i.e. image registration using feature-based methods involving models
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/10—Image acquisition modality
- G06T2207/10116—X-ray image
- G06T2207/10121—Fluoroscopy
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/30—Subject of image; Context of image processing
- G06T2207/30004—Biomedical image processing
- G06T2207/30101—Blood vessel; Artery; Vein; Vascular
Abstract
The invention discloses an optical coherence angiography imaging method based on machine learning. According to the OCT three-dimensional structure image acquisition method, OCT three-dimensional structure images of a sample acquired by OCTA equipment are utilized to generate an original data set required by network model training, a whole group of OCT structure images with poor registration effect are removed, an OCTA algorithm is adopted to carry out contrast imaging to generate a training data set, a machine learning network model is built, and the machine learning network model is trained, so that OCTA contrast is carried out through the machine learning network model; the invention can play a great role in the OCTA field, can generate an angiogram with higher signal-to-noise ratio and better vascular connectivity, and can inhibit the common speckle effect in OCT images to a great extent; the label image is automatically generated by an algorithm, so that the applicability of the method is enlarged without being influenced by own system errors caused by different systems; the imaging can be performed with less detection power to reduce the damage or the amount of data required for imaging can be reduced during imaging, enabling faster completion of the scan.
Description
Technical Field
The invention relates to an optical coherence angiography imaging technology, in particular to an optical coherence angiography imaging method based on machine learning.
Background
Optical coherence tomography (Optical Coherent Tomography, OCT) is a high resolution, non-contact, fast three-dimensional imaging technique. The method utilizes the coherent principle of scattered light in biological tissues, and the signal contrast comes from the difference of light scattering capacity of different biological tissues. The OCT technology combines a semiconductor technology and an ultrafast laser technology, and a broadband light source, a Michelson interferometer, a photoelectric detector and other core components are utilized to obtain a back scattering signal of the biological tissue, so that a real-time micron-sized tomographic image of the biological tissue can be finally obtained through digital signal processing of a computer. Therefore, OCT technology has long become one of the important means for anatomical image diagnosis, which plays an important role not only in ophthalmic clinical examinations but also in fields such as dermatology, gastroenterology, cardiology, and neurology.
With the development of scientific technology, OCT technology has undergone multiple major breakthroughs and developments in hardware and software in the last 30 years, with faster imaging speeds and higher system sensitivity. In particular, after 2002, as the frequency domain OCT technology matures, the OCT technology has gained attention and application in various fields.
In 1991, huang et al, the university of american college of hemp and technology built a first OCT prototype with a longitudinal resolution of 15 μm, and published a first in-vitro human eye retina OCT scan image and corresponding tissue slice images in Science journal, verifying the feasibility of the OCT system. Wojtkowski et al in 2002 obtained the first in-vivo human eye retina image in the world based on the frequency domain OCT technique, johannes and Leitgeb successively compared the frequency domain OCT in theory and experiment with each parameter in time domain OCT, and proved that the frequency domain OCT has higher sensitivity and faster imaging speed. Since then, frequency domain OCT has gradually replaced time domain OCT, and has gained widespread attention and application.
Optical coherence angiography imaging (Optical Coherent Tomography Angiography, OCTA) is a new type of non-invasive vascular imaging technique that has emerged in recent years. During specific imaging, the signal light scans the sample through the galvanometer system, the scanning area is generally rectangular and is divided into a fast axis direction and a slow axis direction, and during scanning, the signal light is continuously scanned repeatedly for a plurality of times (generally 4 times) in the fast axis direction, so that OCT signals of the same position at different moments are recorded, then tissue information is removed through algorithm processing, blood flow signals are extracted, and angiography images are generated. The method skillfully uses flowing red blood cells as contrast agents, namely when the red blood cells continuously flow in blood vessels, OCT signals in the blood vessels continuously change, and the OCT signals are distinguished from stable signals of static tissues.
The current OCTA imaging algorithms are mainly classified into three types of imaging algorithms based on phase change, amplitude change and phase and amplitude combined change according to the source of blood vessel information. The essence is to compare OCT signals at different moments at the same position in an analytic calculation mode. However, these methods often only use a part of information in OCT signals, resulting in problems such as low signal-to-noise ratio of contrast images and serious speckles. The main means for solving the problem is to increase the scanning times of the same position and strengthen the intensity of blood vessel signals, and the method can lead to overlong scanning time, and sample tremble can generate artifacts, such as tremble and breathing of eyes of patients during ophthalmic examination. In addition, long-term laser irradiation can also cause damage to biological tissue.
Disclosure of Invention
Aiming at the problems in the prior art, the invention provides an optical coherence angiography imaging method based on machine learning.
The invention discloses an optical coherence angiography imaging method based on machine learning, which comprises the following steps:
1) Generating an original data set:
generating an original data set required by network model training by utilizing OCTA equipment to acquire OCT three-dimensional structure images of a sample, wherein the original data set comprises j multiplied by k groups of OCT structure image sequences, each group of OCT structure image sequences comprises i OCT structure images with two-dimensional cross sections (B-Scan), k is the number of samples, j is the number of slow-axis scanning positions of each sample, i is the scanning times of the same slow-axis scanning position of the same sample, i is a natural number of > 4, j is a natural number of > 50, and k is a natural number of > 5;
2) Data screening:
registering the i B-Scan surface OCT structural images in the same group of OCT structural images by adopting a rigid registration algorithm, and utilizing after registrationThe correlation algorithm calculates the registration accuracy, eliminates the whole group of OCT structure images with poor registration effect, reserves n groups of OCT structure images after screening, wherein n is a natural number
3) Generating a training data set:
carrying out contrast imaging by using n groups of screened OCT structure images obtained in the step 2) and adopting an OCTA algorithm, wherein each group of OCT structure images can obtain an OCTA contrast image of a B-Scan surface, which is called a label image; taking out m B-Scan face OCT structure images, called input data, from i B-Scan face OCT structure images of a set of OCT structure images corresponding to each of the label images, paired with the label images, the input data and the label images constituting a training data set required for training of the network model, wherein m=2, 3 or 4;
4) Establishing a machine learning network model:
constructing a machine learning network model and setting super parameters of the machine learning network model; and dividing the training data set into n 1 Group training set and n 2 Group test set, training set and test set are independent, n 1 And n 2 Are respectively natural numbers, and
5) Training a machine learning network model:
using the machine learning network model established in the step 4), taking the input data in the training data set as the input of the machine learning network model, wherein n is as follows 1 The group training set is used for training a machine learning network model and is represented by n 2 The group test set is used for checking the performance of the machine learning network model; in the training process, the training set is trained for a plurality of rounds by dividing a plurality of batches and repeatedly inputting the training set into the machine learning network model, and meanwhile, the difference between the output image and the label image of the machine learning network model is judged or calculated as a training errorAfter the training of each batch is finished, performing performance test on the machine learning network model by using a test set, and when the performance test index training of the standby machine learning network model tends to be stable, considering that the training of the machine learning network model is finished, and storing the machine learning network model after the training is finished;
6) Performing OCTA contrast by a machine learning network model:
and (3) taking the OCT structure image of the sample acquired by the OCTA equipment as input by using the trained machine learning network model, and taking the output image as an OCTA contrast image.
In step 1), when the OCTA device collects samples, scanning a slow axis scanning position of a scanning position of one sample once to obtain an OCT structure image of a B-Scan plane, scanning the same slow axis scanning position i times, each sample having j slow axis scanning positions, and k samples in total, thereby obtaining an OCT structure image of an i×j×k B-Scan plane, dividing the OCT structure image of the i×j×k B-Scan plane into j×k groups of OCT structure images, each group of OCT structure images including the OCT structure images of the i B-Scan plane, that is, each slow axis scanning position corresponding to one group of OCT structure images, thereby obtaining an original data set.
In the step 2), for the OCT structural images after j×k groups of registration, the first n groups of OCT structural images with higher registration accuracy are retained, the rest consider that the registration effect is poor, and the OCT structural images are removed.
In step 4), the machine learning network model adopts a deep convolutional neural network CNN, a generative countermeasure network GAN or a cyclic neural network RNN; super parameters include network layer number, convolution kernel, learning rate, parameter initialization, training round number and batch size.
In step 5), the training error is minimized using one of a random gradient descent (Stochastic gradient descent, SGD), an adaptive moment estimation optimization algorithm (Adaptive Moment Estimation, adam), and a Momentum algorithm (Momentum) with the mean square error, structural similarity, or peak signal-to-noise ratio between the output image and the label image as a training error and performance test index to train the machine learning network model.
In the step 6), the OCT structure image of the sample is scanned for a plurality of times at the same slow axis scanning position, and the scanning times are less than or equal to 4.
Machine learning is an inevitable product of the development of artificial intelligence research to a certain stage, and has become a core research topic of artificial intelligence. The aim is to let the computer learn knowledge or skills by mimicking the behavior learned by humans and to learn new knowledge continuously to improve performance. Machine learning refers to the subjects of physiology, psychology, cognition and the like, and through the knowledge of self-learning mechanism of human, a calculation model or a cognition model similar to human learning is established, so that various learning theory and learning methods are formed, and a learning system with specific application is established for specific tasks.
Common algorithms in machine learning at present include artificial neural networks, support vector machines, naive bayes, random forests, sparse dictionaries, reinforcement learning, characterization learning, similarity metric learning and the like. With the development of computer hardware, deep learning is gradually developed, and the deep learning is a comprehensive evolution of an artificial neural network. The deep learning expands the depth and width of the artificial neural network, and can infinitely approximate to a more complex nonlinear model, so that objective rules and internal relations hidden in data are learned. In general terms, deep learning algorithms include deep belief networks, deep neural networks, and convolutional neural networks, where the deep belief networks and deep neural networks are very similar in structure. The deep learning network with more image processing is a convolutional neural network and a generative countermeasure network. At present, machine learning is widely applied to the fields of reconstruction, enhancement and segmentation of medical images, but is not yet applied to image reconstruction of OCTA.
The invention has the advantages that:
the invention can play a great role in the OCTA field, and the strong data mining capability of the invention helps OCTA equipment to generate an angiogram with higher signal-to-noise ratio and better vascular connectivity, and inhibits the common speckle effect in OCT images to a great extent; it is worth mentioning that the label image in the invention is automatically generated by an algorithm, unlike the common machine learning application, label data is needed to be obtained through expert labeling, and the applicability of the method is enlarged without being influenced by own systematic errors caused by different systems. In addition, in the same OCTA equipment, in order to obtain the OCTA imaging image with the same level, the invention can use smaller detection power to carry out imaging, reduce the damage of laser to biological tissues (such as ophthalmology), or reduce the data volume required by imaging during imaging, namely reduce the scanning times of the same position, can complete scanning more quickly, and reduce artifacts (such as shaking, breathing and the like of a patient during fundus imaging) caused by overlong scanning time and sample trembling.
Drawings
FIG. 1 is a flow chart of a machine learning based optical coherence angiography imaging method of the present invention;
fig. 2 is an OCTA image obtained according to one embodiment of the machine learning based optical coherence angiography imaging method according to the present invention.
Detailed Description
The invention will be further elucidated by means of specific embodiments in conjunction with the accompanying drawings.
The machine learning-based optical coherence angiography imaging method of the present embodiment, as shown in fig. 1, includes the following steps:
1) Generating an original data set:
generating an original data set required by network model training by utilizing OCTA equipment to acquire OCT three-dimensional structural images of retina, scanning the same slow axis scanning position of the same sample for 50 times, wherein the slow axis scanning position of each sample is 100, and the sample is 30 human eyes, so that the original data set comprises 100X 30 OCT structural image sequences, and each OCT structural image sequence comprises 50 OCT structural images of B-Scan surface;
2) Data screening:
registering 50B-Scan surface OCT structure images in the same group of OCT structure images by adopting a rigid registration algorithm, calculating registration accuracy by utilizing a correlation algorithm after registration, eliminating the whole group of OCT structure images with poor registration effect, and reserving 70 percent, namely 100 multiplied by 30 multiplied by 0.7, of screened OCT structure images;
3) Generating a training data set:
performing contrast imaging by using 100 multiplied by 30 multiplied by 0.7 groups of OCT structure images obtained in the step 2) and adopting an OCTA algorithm, wherein each group of OCT structure images can obtain an OCTA contrast image of a B-Scan surface, which is called a label image; taking out 4B-Scan face OCT structure images from 50B-Scan face OCT structure images of a group of OCT structure images corresponding to each label image, namely input data, matching the label images, and forming a training data set required by training a network model by the input data and the label images;
4) Establishing a machine learning network model:
the machine learning network model adopts a deep convolutional neural network DnCNN to form 20 layers of convolutional layers, wherein the layer 1 uses 64 convolution check input 4 OCT structure images of 3 multiplied by 4 to carry out convolution, 64 feature images are generated, and a ReLU function is used as an activation function; each of the layers 2 to 19 uses 64 convolution cores of 3×3×64 to convolve the feature map of the previous layer, and connects the ReLU function as output after batch normalization; the 20 th layer uses 1 64 convolution cores of 3 x 4 to convolve the feature images of the previous layer, and the output image is the output image of the network; in the network parameters, the learning rate is set to be 0.001, the network parameter initialization uses a Kaiming initialization method, the number of repeated training rounds is 50, the batch size is 16, and the ratio of the training set to the test set is 7:3, a step of;
5) Training a machine learning network model:
using the machine learning network model established in the step 4), taking input data in a training data set as input of the machine learning network model, wherein a 100×30×0.7×0.7 group training set is used for training the machine learning network model, and a 100×30×0.7×0.3 group testing set is used for checking the performance of the machine learning network model; in the training process, the training set trains 50 rounds by inputting the machine learning network model in batches and repeatedly, meanwhile, calculates the mean square error between the output image and the label image of the machine learning network model as a training error, and uses an adaptive moment estimation optimization algorithm (Adam) to minimize the training error so as to train the machine learning network model; after the training of each batch is finished, performing performance test on the machine learning network model by using a test set, and when the performance test index (the mean square error between the output image and the label image) of the standby machine learning network model tends to be stable, considering that the training of the machine learning network model is finished, and storing the machine learning network model after the training is finished;
6) Performing OCTA contrast by a machine learning network model:
and (3) taking OCT structural images of the sample acquired by the OCTA equipment as input by using the trained machine learning network model, wherein the scanning times of the same slow axis scanning position are 4 times, and the output images are OCTA contrast images, as shown in figure 2.
Finally, it should be noted that the examples are disclosed for the purpose of aiding in the further understanding of the present invention, but those skilled in the art will appreciate that: various alternatives and modifications are possible without departing from the spirit and scope of the invention and the appended claims. Therefore, the invention should not be limited to the disclosed embodiments, but rather the scope of the invention is defined by the appended claims.
Claims (6)
1. An optical coherence angiography imaging method based on machine learning, characterized in that the optical coherence angiography imaging method comprises the following steps:
1) Generating an original data set:
OCT three-dimensional structure image of sample acquired by OCTA equipment is utilized to generate an original data set required by network model training, wherein the original data set comprisesEach group of OCT structure image sequences comprises i OCT structure images of a two-dimensional cross section B-Scan, wherein k is the number of samples, j is the number of slow-axis scanning positions of each sample, i is the scanning times of the same slow-axis scanning position of the same sample, i is a natural number > 4, j is a natural number > 50, and k is a natural number > 5;
2) Data screening:
registering the i B-Scan surface OCT structural images in the same group of OCT structural images by adopting a rigid registration algorithm, calculating registration accuracy by utilizing a correlation algorithm after registration, eliminating the whole group of OCT structural images with poor registration effect, reserving n groups of screened OCT structural images, wherein n is a natural number, and≤n≤/>；
3) Generating a training data set:
carrying out contrast imaging by using n groups of screened OCT structure images obtained in the step 2) and adopting an OCTA algorithm, wherein each group of OCT structure images can obtain an OCTA contrast image of a B-Scan surface, which is called a label image; taking out m B-Scan face OCT structure images, called input data, from i B-Scan face OCT structure images of a set of OCT structure images corresponding to each of the label images, paired with the label images, the input data and the label images constituting a training data set required for training of the network model, wherein m=2, 3 or 4;
4) Establishing a machine learning network model:
constructing a machine learning network model and setting super parameters of the machine learning network model; and divide the training data set intoGroup training set and +.>Group test set, training set and test set are independent of each other, < ->And->Are natural numbers respectively, and->+/>=n，/>≥/>；
5) Training a machine learning network model:
using the machine learning network model established in the step 4), taking the input data in the training data set as the input of the machine learning network model, whereinThe group training set is used for training a machine learning network model and is in +.>The group test set is used for checking the performance of the machine learning network model; in the training process, the training set is divided into a plurality of batches and repeatedly input into the machine learning network model for training for a plurality of rounds, meanwhile, the difference between the output image and the label image of the machine learning network model is judged or calculated to serve as a training error to train the machine learning network model, after the training of each batch is finished, the test set is used for carrying out performance test on the machine learning network model, when the performance test index of the standby machine learning network model is trained to be stable, the training of the machine learning network model is considered to be completed, and the machine learning network model with the completed training is saved;
6) Performing OCTA contrast by a machine learning network model:
and (3) taking the OCT structure image of the sample acquired by the OCTA equipment as input by using the trained machine learning network model, and taking the output image as an OCTA contrast image.
2. The optical coherence angiography imaging method of claim 1, wherein, in stepIn step 1), when the OCTA equipment collects samples, scanning one slow axis scanning position of one sample once to obtain an OCT structure image of a B-Scan surface, scanning the same slow axis scanning position i times, wherein each sample has j slow axis scanning positions, and k samples are taken in total, so that i is obtainedOCT structure image of the B-Scan plane will be i +.>OCT structural image of the individual B-Scan plane is divided into +.>And each group of OCT structure images comprises i OCT structure images of a B-Scan plane, namely, each slow axis scanning position corresponds to one group of OCT structure images, so that an original data set is obtained.
3. The optical coherence angiography imaging method of claim 1, wherein in step 2), forThe OCT structure images after the group registration are compared with each other, the top n groups of OCT structure images with higher registration accuracy are reserved, the rest are considered to have poorer registration effect, and the OCT structure images are removed.
4. The optical coherence angiography imaging method of claim 1, wherein in step 4), the machine-learning network model employs a deep convolutional neural network CNN, a generative antagonism network GAN, or a recurrent neural network RNN.
5. The optical coherence angiography imaging method of claim 1, wherein in step 4), the super-parameters include network layer number, convolution kernel, learning rate, parameter initialization, training round number, and lot size.
6. The optical coherence angiography imaging method of claim 1, wherein in step 5), a mean square error, structural similarity, or peak signal-to-noise ratio between the output image and the label image is used as a training error, and one of a random gradient descent, an adaptive moment estimation optimization algorithm, and a momentum algorithm is used to minimize the training error to train the machine-learned network model.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910513946.0A CN112085830B (en) | 2019-06-14 | 2019-06-14 | Optical coherence angiography imaging method based on machine learning |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910513946.0A CN112085830B (en) | 2019-06-14 | 2019-06-14 | Optical coherence angiography imaging method based on machine learning |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN112085830A CN112085830A (en) | 2020-12-15 |
| CN112085830B true CN112085830B (en) | 2024-02-27 |
Family
ID=73733802
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910513946.0A Active CN112085830B (en) | 2019-06-14 | 2019-06-14 | Optical coherence angiography imaging method based on machine learning |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN112085830B (en) |
Families Citing this family (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114209278B (en) * | 2021-12-14 | 2023-08-25 | 复旦大学 | Deep learning skin disease diagnosis system based on optical coherence tomography |
| CN116313115B (en) * | 2023-05-10 | 2023-08-15 | 浙江大学 | Drug action mechanism prediction method based on mitochondrial dynamic phenotype and deep learning |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107865642A (en) * | 2017-09-28 | 2018-04-03 | 温州医科大学 | A kind of glaucoma filtering operation avascular filtering bleb evaluation method based on OCT Angiographies |
| CN108670239A (en) * | 2018-05-22 | 2018-10-19 | 浙江大学 | A kind of the three-dimensional flow imaging method and system in feature based space |
Family Cites Families (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2010138645A2 (en) * | 2009-05-29 | 2010-12-02 | University Of Pittsburgh - Of The Commonwealth System Of Higher Education | Blood vessel segmentation with three-dimensional spectral domain optical coherence tomography |
| JP7220509B2 (en) * | 2017-09-27 | 2023-02-10 | 株式会社トプコン | OPHTHALMIC DEVICE AND OPHTHALMIC IMAGE PROCESSING METHOD |
-
2019
- 2019-06-14 CN CN201910513946.0A patent/CN112085830B/en active Active
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107865642A (en) * | 2017-09-28 | 2018-04-03 | 温州医科大学 | A kind of glaucoma filtering operation avascular filtering bleb evaluation method based on OCT Angiographies |
| CN108670239A (en) * | 2018-05-22 | 2018-10-19 | 浙江大学 | A kind of the three-dimensional flow imaging method and system in feature based space |
Non-Patent Citations (1)
| Title |
|---|
| 高灵敏、高对比度无标记三维光学微血管造影系统与脑科学应用研究;周丽萍;李培;潘聪;郭立;丁志华;李鹏;物理学报;第65卷(第15期);全文 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN112085830A (en) | 2020-12-15 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US20220058803A1 (en) | System for oct image translation, ophthalmic image denoising, and neural network therefor | |
| US20210304363A1 (en) | Image processing apparatus, image processing method and computer-readable medium | |
| US10719932B2 (en) | Identifying suspicious areas in ophthalmic data | |
| CN106028921A (en) | Optical coherence tomography system for health characterization of an eye | |
| CN112085830B (en) | Optical coherence angiography imaging method based on machine learning | |
| CN114390907A (en) | Machine learning method for creating a structure-derived field-of-view prior | |
| CN106028912A (en) | Optical coherence tomography (oct) system with improved motion contrast | |
| US20230196572A1 (en) | Method and system for an end-to-end deep learning based optical coherence tomography (oct) multi retinal layer segmentation | |
| Awasthi et al. | Sinogram super-resolution and denoising convolutional neural network (SRCN) for limited data photoacoustic tomography | |
| Tajmirriahi et al. | A lightweight mimic convolutional auto-encoder for denoising retinal optical coherence tomography images | |
| JP2020039851A (en) | Image processing device, image processing method and program | |
| CN115349137A (en) | Correction of flow projection artifacts in OCTA volumes using neural networks | |
| Li et al. | Deep learning algorithm for generating optical coherence tomography angiography (OCTA) maps of the retinal vasculature | |
| WO2021106967A1 (en) | Ocular fundus image processing device and ocular fundus image processing program | |
| CN115443480A (en) | OCT orthopathology segmentation using channel-encoded slabs | |
| Zaki et al. | Towards automated keratoconus screening approach using lateral segment photographed images | |
| Sultana et al. | RIMNet: image magnification network with residual block for retinal blood vessel segmentation | |
| EP3667618A1 (en) | Deep partial-angle coronary restoration | |
| Viedma et al. | A deep learning multi-capture segmentation modality for retinal OCT imaging | |
| CN115880310B (en) | Retina OCT fault segmentation method, device and equipment | |
| Cisneros-Guzmán et al. | Segmentation of OCT and OCT-A Images using Convolutional Neural Networks | |
| US20230190095A1 (en) | Method and system for choroid-scleral segmentation using deep learning with a choroid-scleral layer model | |
| Zhao et al. | Automatic generation of retinal optical coherence tomography images based on generative adversarial networks | |
| Gende et al. | A new generative approach for optical coherence tomography data scarcity: unpaired mutual conversion between scanning presets | |
| Ma et al. | Deep‐learning segmentation method for optical coherence tomography angiography in ophthalmology |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |